Motion segmentation in videos from time of flight cameras

This paper investigates motion estimation and segmentation of independently moving objects in video sequences from a range camera. Specifically, we present a motion estimation algorithm which is based on fusion of range flow and optical flow constraint equations. The flow fields are used to derive long-term point trajectories. A segmentation technique groups the trajectories according to their motion and depth similarity into spatio-temporal objects. We show results on a real world scene captured with a time of flight camera.